The Venture Codex

Participated · Seed · Jun 2026

Valence AI develops Pulse Emotion models and a patented audio signal processing pipeline that analyze live calls to classify emotional state from speech in real time, converting tone, pacing, and vocal cues into structured data. The company publishes its own foundation models trained on proprietary, demographically and neurotype-diverse speech datasets and reports 92% accuracy on internal benchmarks. Products include emotion-aware IVR, an Agent Assist copilot for live call coaching, emotionally intelligent AI voice agents, and post-call quality assurance; integrations include ElevenLabs and Cartesia. In production deployments Valence reports handle time reductions of 30% and improvements in customer satisfaction scores, close rates, and time to close, and it serves customers such as Harte Hanks, CustomerHD, and BPO Centers. Valence holds two U.S. patents issued in June 2026 covering its core signal processing pipeline and an extension for live haptic feedback, and its offerings are SOC 2 Type 2 and HIPAA compliant. The company plans to expand language coverage and deepen platform integrations as part of its near-term roadmap.

Participated · Equity · Jul 2025

Synfini operates an AI Cloud Foundry platform that merges artificial intelligence, automated synthesis, and iterative molecular design to accelerate discovery and synthesis of small molecules. The platform integrates target selection, compound design, reaction route prediction, and synthesis into a single workflow that shortens iteration cycles and reduces costs. By combining physical and virtual chemistry, Synfini produces novel, optimized compounds more rapidly than traditional early discovery processes. Since spinning out from SRI in 2023, the company has confirmed its approach through early traction and strategic partnerships. Synfini will use recent funding to expand its software and automated lab capabilities, hire additional scientific and engineering staff, and deepen commercial partnerships. The company has raised a total of $53 million to date, including the latest $8.9 million expansion and a mix of venture investments and non-dilutive grants.

Participated · Series A · Mar 2024

Pienso offers a no-code AI suite that enables non-technical teams — researchers, marketers and customer support — to annotate data, train and deploy models without writing code. The platform guides users through labeling for pre-tuned open-source or custom models and can be deployed in the cloud or on-premises. It integrates with enterprise systems via APIs but can also operate without third-party services to keep data within a secure environment. Pricing is a yearly license based on the number of AI models a customer deploys, allowing customers to experiment before making large investments. Customers cited in the article include Sky (using Pienso to analyze customer service calls) and an unnamed U.S. government agency that tested it for monitoring illegal weapons tracking. Pienso was founded in 2016 by Birago Jones and Karthik Dinakar, who developed the core tools during research at MIT’s Media Lab. Pienso is a Brooklyn-based MIT spin-off that helps non-technical domain experts interact with and improve machine learning models. Its core product, the Intelligent Development Environment, is an end-to-end platform that includes Lens, a feature that lets non-technical users directly interact with ML algorithms. The company aims to solve the human-in-the-loop problem by embedding domain expertise into model training so nuance and context are preserved. CEO Birago Jones frames Pienso as a way to capture expert judgment from researchers, analysts and other knowledge workers who lack ML expertise. Pienso announced a $2.1 million seed round led by Eniac Ventures with participation from SoftTech VC, Indicator Ventures and E14 Fund. It plans to use the funding to grow its small team and start expanding its customer base.

Participated · Seed · Nov 2023

Zephr is building a "networked GPS" that uses groups of smartphones to collect raw satellite measurements, computes error corrections on a server, and returns those corrections to improve device positioning. The system directs nearby phones to ping GPS measurements and aggregates them to produce corrections similar to differential base stations. The company emerged from stealth and has raised a $3.5 million seed round. Zephr validated its approach with simulations and testing at SRI International and has run field tests using real phones and RTK‑GPS systems for ground truth. It plans pilots via a demo app that compares standard phone GPS to the corrected output and to customize models for customers. The team (six full‑time, two part‑time) includes founders Sean Gorman, Pramukta Rao, Kostas Stamatiou and Scotty Nelson, with prior experience at Snap, Violet Labs, BlueSky Resources and Twitter/X. Zephr has seen interest from rideshare, location‑based gaming and adtech verticals and is positioning the technology for longer‑term markets such as autonomy, robotics and augmented reality.

Participated · Seed · Jul 2022

Lisa Health, spun out of SRI International in 2020, develops Midday, an AI-driven app to support women through menopause. The app delivers personalized insights, tracks a user’s stage in the menopause journey, and recommends holistic interventions and symptom treatments. Midday offers users the ability to request virtual care appointments with Mayo Clinic clinicians and provides guidance on hormone therapy eligibility. Lisa Health developed the app under a health license agreement with Mayo Clinic, which provided clinical expertise and continues to advise on evidence-backed approaches. The company is working on a premium version that will include novel technology for hot flash detection and other menopause biomarkers. Midday is available on the app store with a 30-day free trial, then a $29.99 quarterly or $79.99 annual subscription. Lisa Health positions itself against competitors such as Evernow, Vira Health, and Alloy while emphasizing an AI- and physiology-based approach.